Prediction of deflection from flatness and a vertical position with the use of neural networks

Journal Title: Civil and Environmental Engineering Reports - Year 2012, Vol 1, Issue 9

Abstract

The paper presents an attempt to apply unidirectional multilayer neural networks in the prediction of the deflections from flatness and from a vertical position of building walls, on an example of periodic measurements in the church of the Blessed Virgin Mary in Toruń. The applied methods of artificial intelligence in a form of sigmoid neural networks were taught with the use of the backpropagation method, which bases on the gradient methods described in optimization theories. The prognosis of the values of the deflections from flatness and from vertical position was carried out for a single measurement epoch on the basis of ten periodic measurements performed at several-year intervals.<br/><br/>

Authors and Affiliations

M. Mrówczyńska

Keywords

Related Articles

EVALUATION OF DYNAMIC CHARACTERISTICS OF THE FOOTBRIDGE WITH INTEGRAL ABUTMENTS

The paper presents the results of dynamic field tests and numerical analysis of the footbridge designed as a three-span composite structure with integral abutments. The adopted design solution which has allowed to achiev...

Permeability, tortuosity, and attenuation of waves in porous materials

The paper contains a consistent presentation of the construction of a linear poroelastic model and its applications in the theory of acoustic waves. The main purpose of this presentation is the discussion of material par...

The sozological and hygrographic maps of Poland as a basis for the analysis of environmental changes in mining areas

In the forthcoming decades large resources of lignite are going to play a role of energy security stabiliser in Poland. This proves the significance of the problems of mining areas in geographical space and entails an in...

COMPARISON OF THE CALCULATIONS RESULTS OF HEAT EXCHANGE BETWEEN A SINGLE-FAMILY BUILDING AND THE GROUND OBTAINED WITH THE QUASI-STATIONARY AND 3-D TRANSIENT MODELS. PART 2: INTERMITTENT AND REDUCED HEATING MODE

The paper provides comparative results of calculations of heat exchange between ground and typical residential buildings using simplified (quasi-stationary) and more accurate (transient, three-dimensional) methods. Such...

ALUMINUM RECOVERY FROM MULTIMATERIAL TETRA-PAK WASTE PYROLYSIS

A constant growth of the multimaterial waste production can be observed in the recent years. The multimaterial waste that contain aluminum are especially hard to process due to the fact that multiple layers of various ma...

Download PDF file
  • EP ID EP346695
  • DOI -
  • Views 56
  • Downloads 0

How To Cite

M. Mrówczyńska (2012). Prediction of deflection from flatness and a vertical position with the use of neural networks. Civil and Environmental Engineering Reports, 1(9), 73-81. https://europub.co.uk/articles/-A-346695